Conversational Case-Based Planning for Agent Team Coordination

Abstract

This paper describes a prototype in which a conversational case-based reasoner, NaCoDAE, was agentified and inserted in the RET- SINA multi-agent system. Its task was to determine agent roles within a heterogeneous society of agents, where the agents may use capability- based or team-oriented agent coordination strategies. There were three reasons for assigning this task to NaCoDAE: (1) to relieve the agents of the overhead of determining, for themselves, if they should be involved in the task, or not; (2) to convert seemingly unrelated data into contextually relevant knowledge — as a case-based reasoning system, NaCo-DAE is particularly suited for applying apparently incoherent data to a wide variety of domain-specific situations; and (3) as a conversational CBR system, to both unobtrusively listen to human statements and to proactively dialogue with other agents in a more goal-directed approach to gathering relevant information. The cases maintained by NaCoDAE have question and answer components, which were originally intended to maintain the textual representations of questions and answers for humans. By associating agent capability descriptions and queries with the case questions, NaCoDAE also assumed the team role of a capability- based coordinator. By encoding fragments of HTN plan objectives in its case actions, we were able to convert NaCoDAE into a conversational case-based planner that served compositionally-generated HTN plan objectives, already populated with situation-relevant knowledge, for use by the RETSINA team-oriented agents.

The authors are grateful to the Naval Research Labs for providing the sources to Na- CoDAE. Matthew W. Easterday made a significant contribution to this project by adapting NaCoDAE to operate in an agent context. Many thanks to Alex Rudnicky for allowing us to agentify Sphinx and for providing us with technical support. On a personal note, Joseph Giampapa would like to thank David Aha for his encouragement and helpful suggestions. This research was sponsored in part by the Office of Naval Research Grant N-00014-96-16-1-1222 and by DARPA Grant F-30602-98- 2-0138.